Research On The Multi-sensor Fusion Sensing System For The Lower Exoskeleton Robot

Power-assisted lower limb exoskeleton robot is originally used to enhance the physical fitness of soldiers,improve individual combat capability,widely used in rehabilitation,earthquake relief and climbing power and other occasions,many researchers have been engaged in the exoskeleton robot research.The main study is divided into three aspects:bionic design,precise perception and compliant control.The precise perception of the wearer's intention is the premise of compliant control and the key to realizing human-machine coupling.Therefore the sensing system of the self-developed lower limb exoskeleton robot in the thesis has been studied in the following parts.Chapter 1:Introduction.The review of the sensing system of self-developed lower limb exoskeleton robot.The research background and the domestic and foreign research status are introduced,the goals and significances and research contents of the study are arranged.Chapter 2:Analysis of sensing parameters of the self-developed lower exoskeleton robot.The system architecture of the self-developed lower limb exoskeleton robot is introduced systematically and the lower limb dynamic model is analyzed.Based on this,the cooperative control method of the lower limb exoskeleton robot is studied.The support phase adopts position control and the swing phase adopts model-based interactive force control method.Finally,the input of the hybrid control method is determined as the perceived quantity of the sensing system.Chapter 3:Acquirement and analysis of sensing parameters.Based on the full understanding of the lower limb kinematics theory,the gait information acquisition system is built for meeting with the control demand of the self-developed exoskeleton robot.The data of IMU and FSR is processed by the variable-gain Kalman data fusion algorithm and the Savitzky-Golay filter algorithm respectively,by this way,the sensing quantities can be acquired directly at the sensor.Chapter 4:Research on the sensing and recognition algorithm of the self-developed lower exoskeleton robot.Mainly for algorithm research for the sensing quantity can not be directly obtained from the sensor.The fuzzy-proportional gait recognition algorithm based on plantar pressure is proposed to identify the four gait phases of the heel strike,load response,pre-swing and swing.The implementation of the algorithm is discussed in detail.Chapter 5:Experimental study on the sensing system of the self-developed lower exoskeleton robot.The experiments are carried out for the algorithm of attitude-angle calculating,the method of man-machine interactive and the algorithm of fuzzy-proportional gait recognition based on plantar pressure.The results show that the proposed algorithm of attitude-angle calculating has high precision and reliability.In addition,the feasibility of the man-machine interactive method is verified by the human-computer experiment.The algorithm of fuzzy-proportional gait recognition is able to recognize the gait phase with an average accuracy of 99%or more,at the same time,the rapidity of algorithm is verified by the foot switch as a reference system.Chapter 6:Summary and prospect.The research content of this paper is summarized,in addition,the existing problems and future research directions are outlined.